Data-Driven State Prediction and Sensor Fault Diagnosis for Multi-Agent Systems with Application to a Twin Rotational Inverted Pendulum

Xin Lu,Bowen Li,Jie Zhong,Xiaoxu Liu

doi:10.3390/pr9091505

Abstract

When a multi-agent system is subjected to faults, it is necessary to detect and classify the faults in time. This paper is motivated to propose a data-driven state prediction and sensor fault classification technique. Firstly, neural network-based state prediction model is trained through historical input and output data of the system. Then, the trained model is implemented to the real-time system to predict the system state and output in absence of fault. By comparing the predicted healthy output and the measured output, which can be abnormal in case of sensor faults, a residual signal can be generated. When a sensor fault occurs, the residual signal exceeds the threshold, a fault classification technique is triggered to distinguish fault types. Finally, the designed data-driven state prediction and fault classification algorithms are verified through a twin rotational inverted pendulum system with leader-follower mechanism.

Highlights

Monitoring the condition of complex systems in real-time can save valuable time and cost to maintain the system
Knowledge-based method is named data-driven method, where a fault diagnosis model is built through historical data rather than precise mathematical model
The communication among agents are internal in the agents but unknown in state prediction and fault classification, which implies that the designed state prediction and fault diagnosis techniques are fully distributed

Summary

Introduction

Monitoring the condition of complex systems in real-time can save valuable time and cost to maintain the system. Event-triggered fault diagnosis methods have been developed in [21,22], where the mathematical model of the system is assumed to be known. It is motivated to develop event-triggered data driven fault diagnosis for MAS with unknown mathematical model and unknown communication. If the residual exceeds the threshold, it triggers a fault classification training process to identify and locate the fault This residual-triggered fault diagnosis method does not depend on a mathematical model and communication information. The communication among agents are internal in the agents but unknown (not available) in state prediction and fault classification, which implies that the designed state prediction and fault diagnosis techniques are fully distributed.

Data-Driven State Prediction for Multi-Agent System

Neural Network-Based State Prediction

Fault Classification

Conclusions